How to add alternative text for an image in Tagged PDF?

Tags: PDF/UAtagged pdfimagesiText 5

I know that iText can generate tagged PDF documents from scratch, but is it possible to insert alternative text for images in an existing tagged PDF without changing anything else? I need to implement this feature in a program without using GUI applications such as Adobe Acrobat Pro.

Posted on StackOverflow on Dec 2, 2015 by tsforsure

Please take a look at the AddAltTags example.

In this example, we take a PDF with images of a fox and a dog where the Alt keys are missing: no_alt_attribute.pdf

Structure element without /Alt key
Structure element without /Alt key

Code can't recognize a fox or a dog, so we create a new document with Alt attributes saying "Figure without an Alt description": added_alt_attributes.pdf)

Structure element with /Alt key
Structure element with /Alt key

We add this description by walking through the structure tree, looking for structural elements marked as /Figure elements:

public void manipulatePdf(String src, String dest)
    throws IOException, DocumentException {
    PdfReader reader = new PdfReader(src);
    PdfDictionary catalog = reader.getCatalog();
    PdfDictionary structTreeRoot =
        catalog.getAsDict(PdfName.STRUCTTREEROOT);
    manipulate(structTreeRoot);
    PdfStamper stamper = new PdfStamper(
        reader, new FileOutputStream(dest));
    stamper.close();
}

public void manipulate(PdfDictionary element) {
    if (element == null)
        return;
    if (PdfName.FIGURE.equals(element.get(PdfName.S))) {
        element.put(PdfName.ALT,
            new PdfString("Figure without an Alt description"));
    }
    PdfArray kids = element.getAsArray(PdfName.K);
    if (kids == null) return;
    for (int i = 0; i < kids.size(); i++)
        manipulate(kids.getAsDict(i));
}

You can easily port this Java example to C#:

  • Get the root dictionary from the PdfReader object,

  • Get the root of the structure tree (a dictionary),

  • Loop over all the kids of every branch of that tree,

  • When a lead is a figure, add an /Alt entry.

Once this is done, use PdfStamper to save the altered file.